On 16 December 2024, Justice Kennett of the Federal Court of Australia (New South Wales Registry, File No. NSD 217 of 2020) delivered judgment in Commissioner of Taxation v Bakarich (No 2) [2024] FCA 1448, imposing aggregate civil pecuniary penalties of AU$13.6 million against four respondents for promoting unlawful tax exploitation schemes centred on the Australian Research and Development Tax Incentive (R&DTI)1. While a substantial body of Australian Administrative Appeals Tribunal (AAT)—now the Administrative Review Tribunal (ART)—and Federal Court jurisprudence has historically examined R&DTI eligibility disputes directly between claimant taxpayers and either Industry Innovation and Science Australia (supported by AusIndustry within the Department of Industry, Science and Resources (DISR)) or the Commissioner of Taxation, Bakarich (No 2) represents a landmark enforcement action applying the civil promoter penalty regime in Division 290 of Schedule 1 to the Taxation Administration Act 1953 (Cth) (TAA 1953) directly to R&D tax advisory practices1.
The decision demonstrates the dual-pronged enforcement architecture governing Australia’s R&DTI programme4. When an R&D advisory firm fabricates, exaggerates, or negligently rubber-stamps claims under Division 355 of the Income Tax Assessment Act 1997 (Cth) (ITAA 1997), the legal failure triggers two parallel statutory consequences4. First, the underlying corporate taxpayers suffer disallowance of their notional R&D deductions and refundable tax offsets, accompanied by tax shortfall liabilities, shortfall interest charges, and administrative penalties of up to 75% of the shortfall4. Second, the intermediaries, corporate advisory vehicles, controlling directors, and registered tax agents who marketed and executed those claims face severe personal and corporate civil penalties under Division 290 of Schedule 1 to the TAA 19534.
| Case Attribute | Formal Details and Litigation Parameters |
|---|---|
| Case Citation | Commissioner of Taxation v Bakarich (No 2) [2024] FCA 1448 (related earlier penalty ruling cited as Commissioner of Taxation v Bakarich (Penalty) [2022] FCA 1032)3 |
| Court & Registry | Federal Court of Australia, New South Wales Registry (File No. NSD 217 of 2020)2 |
| Presiding Judge | The Honourable Justice Kennett3 |
| Date of Judgment | 16 December 20243 |
| Applicant | Commissioner of Taxation (represented by G O’Mahoney, K Petch, and A Reid, instructed by the Australian Government Solicitor)3 |
| Respondents | 1. Julian Bakarich (First Respondent, sole director of TDC and co-director of TDA) 2. The Dream Consortium Pty Ltd (in liq) (ACN 168 575 666) (Second Respondent) 3. The Dream Accountants Pty Ltd (in liq) (Third Respondent) 4. Thi Cam Tu Nguyen (Fourth Respondent, co-director of TDA and former registered tax agent)3 |
| Primary Statutory Provisions | Sections 290-5, 290-50, 290-55, 290-60, 290-65 and 290-200 of Schedule 1 to the Taxation Administration Act 1953 (Cth); Division 355 of the Income Tax Assessment Act 1997 (Cth); s 191 of the Evidence Act 1995 (Cth); s 21 of the Federal Court of Australia Act 1976 (Cth); s 500(2) of the Corporations Act 2001 (Cth); s 4AA of the Crimes Act 1914 (Cth)3 |
| Final Orders & Holding | Declarations of contravention of s 290-50(1) in respect of 12 distinct tax exploitation schemes involving 11 taxpayers (2015 and 2016 income years) against Mr Bakarich, TDC and TDA, and in respect of 3 of those schemes (2015 income year) against Ms Nguyen; aggregate civil penalties of AU$13.6m (AU$4.5m against Mr Bakarich; AU$4.5m against TDC; AU$4.5m against TDA; AU$100,000 against Ms Nguyen); TDC and TDA ordered to pay the Commissioner’s costs, otherwise no order as to costs3 |
The factual matrix of Bakarich (No 2) originated on 16 March 2014, when the Second Respondent, The Dream Consortium Pty Ltd (TDC), was incorporated, with the First Respondent, Mr Julian Bakarich—a Sydney-based business coach—serving as its sole director, company secretary, and sole shareholder until 20 August 2019, when he was declared bankrupt3. Mr Bakarich had final responsibility for and control over the work of TDC during the relevant period, in which TDC marketed R&D tax incentive services to clients3. In around December 2014, Mr Bakarich and the Fourth Respondent, Ms Thi Cam Tu Nguyen, resolved to carry on business together through an accounting practice to be run by Ms Nguyen in association with TDC, providing accounting and finance functions to small businesses and trading as The Dream Accountants; after an initial period in which the necessary licences and registrations were obtained (including Ms Nguyen becoming registered as a tax agent), the Third Respondent, The Dream Accountants Pty Ltd (TDA), was incorporated on 29 July 20153. Mr Bakarich and Ms Nguyen were the directors of TDA (Ms Nguyen also being its secretary), and TDA’s issued shares were owned 50% by Ms Nguyen and 50% by TDC3.
Between around 2014 and October 2016, the respondents operated a highly systematised, volume-driven advisory model designed to generate refundable R&D tax offsets for small-to-medium enterprise (SME) clients3. The operational mechanics proceeded through five interconnected stages1. First, TDC obtained clients through word of mouth, referral partners and published marketing material advertising R&D tax incentive services, and at initial meetings Mr Bakarich and/or other TDC employees promoted TDC as R&D specialists with all the necessary approvals and accreditations and extensive experience in making R&D claims, advised prospective clients that they were eligible, and explained that TDC would charge a fee based on a percentage of any R&D tax offset obtained3. Second, clients were asked to complete an online questionnaire describing their business operations and finances and how their company was being innovative in its business methods3. Third, acting under Mr Bakarich’s direct supervision and operational control, TDC used the questionnaire responses to prepare and lodge applications to register purported R&D activities with AusIndustry, within a culture that had complete disregard as to whether clients’ activities even qualified for the R&D tax incentive3. Fourth, TDA prepared an R&D Tax Incentive Schedule for each client by apportioning line items in the taxpayer’s profit and loss statement according to an estimate of how much those items were devoted to R&D activities—Mr Bakarich having instructed employees to maximise the value of claims regardless of their legitimacy—and the schedule was then either provided to the client’s own accountant for inclusion in the income tax return or the return was lodged by TDA on the taxpayer’s behalf3. Fifth, the AusIndustry applications listed a TDC or TDA employee as the contact person and indicated that Ms Nguyen, as a registered tax agent, had provided advice in relation to the application, although no such advice was actually provided to the client; Ms Nguyen’s details were recorded in the tax agent section of the application forms after she reviewed the applications, and she also reviewed the information forming the clients’ R&D claims and prepared R&D Schedules, but did not conduct basic and prudent checks and controls3.
To monetise this structure, TDC entered into contingency-fee engagements with its clients, charging a percentage-based commission (typically 15%) calculated directly against the quantum of the R&D tax refund received by the taxpayer, invoicing the client once the tax return had been lodged and then paying a fee to TDA for its work on the claim3. In the specific Federal Court proceedings before Justice Kennett, the Commissioner prosecuted 12 schemes involving 11 taxpayers executed across the 2015 and 2016 income years, involving corporate claimants such as Basebuild Pty Ltd (Basebuild) and Precision Air Electrical Pty Ltd (PAE)3. Had these 12 schemes gone undetected, the scheme participants would have been able to claim around AU$7 million in refundable tax offsets to which they were not entitled, and approximately AU$591,000 was charged to clients for them (AU$590,808.03 receivable by TDC, of which AU$438,691.72 was actually received)3. More broadly, a joint investigation by the ATO and DISR identified approximately 138 taxpayers linked to suspicious R&DTI claims facilitated by the respondents6. When the ATO subsequently audited these claimants and reversed the invalid offsets, the affected client companies were assessed for more than AU$14.5 million in tax shortfalls, shortfall interest charges, and administrative culpability penalties6. Among the 11 taxpayers in the proceeding, Basebuild, Implemented Portfolios Pty Ltd, PAE and Precision Air Pty Ltd suffered administrative penalties and interest charges as a direct consequence of the wrongful claims, and Precision Air Pty Ltd—unable to pay its tax debt and debts owed to other creditors—was placed in liquidation in August 2018, while officers and senior managers of the taxpayers suffered significant personal distress3. Furthermore, Ms Nguyen was deregistered as a tax agent by the Tax Practitioners Board (TPB) in April 20196.
The Commissioner of Taxation instituted civil penalty proceedings in the Federal Court of Australia alleging that Mr Bakarich, TDC, TDA, and Ms Nguyen had each contravened s 290-50(1) of Schedule 1 to the TAA 1953, which mandates that an entity must not engage in conduct that results in that or another entity being a promoter of a tax exploitation scheme1. To establish liability, the Commissioner had to bring each respondent’s conduct within the promoter penalty provisions of the TAA 1953 by reference to R&D claims made under Division 355 of the ITAA 1997; in the event, the contraventions were admitted by Mr Bakarich and Ms Nguyen, and were found against TDC and TDA in the terms of the facts set out in Mr Bakarich’s statement of agreed facts3.
Under s 290-65(1) of Schedule 1 to the TAA 1953, an arrangement constitutes a “tax exploitation scheme” if two cumulative conditions are met at the time of the promotional conduct4. First, under s 290-65(1)(a), having regard to all relevant matters, it must be reasonable to conclude that an entity entering into or carrying out the scheme did so with the sole or dominant purpose of that entity (or another entity) getting a “scheme benefit” from the scheme4. Pursuant to s 284-150(1) of Schedule 1 to the TAA 1953, a “scheme benefit” encompasses either a decrease in a tax-related liability or an increase in an amount that the Commissioner must pay or credit under a taxation law, which directly captures a refundable R&D tax offset claimed under s 355-100 of the ITAA 19974. Second, under s 290-65(1)(b), it must be not reasonably arguable that the scheme benefit is available at law4. Under s 284-15(1) of Schedule 1 to the TAA 1953, a matter is “reasonably arguable” only if, having regard to the relevant authorities (including the ITAA 1997, the IR&D Act, extrinsic parliamentary materials, and judicial decisions), it would be concluded in the circumstances that what is argued for is about as likely to be correct as incorrect, or is more likely to be correct than incorrect2.
Furthermore, to establish that each respondent was a “promoter” under s 290-60(1) of Schedule 1 to the TAA 1953, the Commissioner had to prove three statutory elements: first, that the entity marketed the scheme or otherwise encouraged the growth of the scheme or interest in it; second, that the entity or an associate received (directly or indirectly) consideration in respect of that marketing or encouragement; and third, that having regard to all relevant matters, including the degree of involvement of the entity in the activities, the entity had a substantial role in respect of that marketing or encouragement1. While s 290-60(2) carves out entities that merely provide independent advice, s 290-60(3) protects employees with minor involvement, and s 290-55 provides exceptions for reasonable mistake of fact or reasonable precautions and due diligence, the Court was not required to determine whether any of these statutory shields applied, because the contraventions were admitted (or, for TDC and TDA, found on the agreed facts); the only aspect of s 290-55 addressed by Justice Kennett was that, because Mr Bakarich admitted that his conduct involved tax evasion, s 290-55(6) displaced the time limit for commencing proceedings in s 290-55(4)3.
The judgment did not itself analyse whether the underlying claims satisfied Division 355 of the ITAA 1997: it was common ground, and admitted by Mr Bakarich and Ms Nguyen, that the schemes involved R&D claims under Division 355 where it was not reasonably arguable that those claims were available at law3. Under the R&DTI regime, an R&D entity is only entitled to a tax offset under s 355-100 for notional deductions arising under s 355-205 (expenditure on R&D activities) or s 355-305 (decline in value of R&D depreciating assets) where the expenditure is actually incurred on registered “core R&D activities” (s 355-25) or “supporting R&D activities” (s 355-30)2. The admitted facts revealed a culture at TDC and TDA that had no regard to whether a client’s claim was reasonably arguable—complete disregard as to whether clients’ activities even qualified for the incentive, R&D Schedules prepared by estimating apportionments of profit and loss line items, and instructions to employees to maximise claims regardless of their legitimacy—and Mr Bakarich admitted that his conduct in respect of each scheme involved tax evasion3.
Mr Bakarich and Ms Nguyen entered into statements of agreed facts with the Commissioner pursuant to s 191 of the Evidence Act 1995 (Cth)—the Nguyen statement being filed on 16 June 2022 following an agreement reached on 15 June 2022, and the Bakarich statement being filed on 3 October 2023 after an in-principle agreement was reached early on the morning of 18 September 2023, the date the three-week trial was set to commence—formally admitting to contraventions of s 290-50(1) of Schedule 1 to the TAA 1953—with Mr Bakarich admitting to promoting all 12 prosecuted tax exploitation schemes and Ms Nguyen admitting to promoting 3 of those schemes—and reaching agreement with the Commissioner on proposed penalty quantum3. TDC and TDA had entered external administration on 27 January 2017, with their liquidations on hold pending the outcome of the proceeding, and they took no active role in the proceedings3.
In Bakarich (No 2), Justice Kennett independently scrutinized the evidentiary record and the matters relevant under s 290-50(5) (measured against the maximum penalties fixed by s 290-50(4)) to determine whether the declarations of contravention should be made and whether the proposed civil penalties were appropriate, emphasising that the Court should not act as a mere “rubber stamp” for agreed penalties3. Justice Kennett accepted that each of the 12 schemes was a separate and distinct set of conduct, so that the contraventions did not arise from a single “course of conduct”, with the possible exception of the two schemes promoted to Precision Air Pty Ltd in consecutive income years, for which the parties factored a minor discount into the proposed penalties3. Each of the R&D claims made with the ATO was idiosyncratic, involving discrete events, interactions and communications between Mr Bakarich and the scheme participants3.
In determining the pecuniary penalties, Justice Kennett applied settled Australian civil penalty jurisprudence, specifically citing the High Court of Australia’s decision in Australian Building and Construction Commissioner v Pattinson [2022] HCA 13; 274 CLR 450, alongside Federal Court authorities in Commissioner of Taxation v Bogiatto (No 2) [2021] FCA 98, Australian Competition and Consumer Commission v Geowash Pty Ltd (No 4) [2020] FCA 23, and NW Frozen Foods Pty Ltd v Australian Competition and Consumer Commission (1996) 71 FCR 2853. Justice Kennett underscored that deterrence is the factor to be given the greatest weight, the purpose of a civil penalty being primarily, if not solely, the promotion of compliance through deterrence: the penalty must put a price on contravention sufficiently high to deter repetition by the contravener and by others, and must not be seen as an “acceptable cost of doing business”, while being no greater than necessary to achieve deterrence and not so high as to be oppressive3. His Honour recorded the parties’ submissions, reflected in Mr Bakarich’s admissions, that there was a complete disregard as to whether clients’ claims were reasonably arguable, that the success of the schemes relied on the Commissioner never discovering that the taxpayers were not entitled to the offsets claimed, and that the schemes were inherently difficult to detect because there was no link on the face of the income tax returns between the claims and Mr Bakarich, whose connection was only discovered through a significant forensic investigation3. Applying the course of conduct and totality principles, and satisfied that the agreed penalties for Mr Bakarich and Ms Nguyen and the penalties proposed by the Commissioner for TDC and TDA were appropriate, Justice Kennett imposed civil penalties totalling AU$13.6 million, noting that the Commissioner did not seek leave to enforce any penalty order against TDC or TDA3.
| Respondent Entity / Individual | Role in the Tax Exploitation Schemes | Admitted / Proven Contraventions | Statutory Maximum Basis (Pre-July 2024 Regime) | Final Penalty Imposed by Federal Court | Primary Judicial Rationale (Kennett J) |
|---|---|---|---|---|---|
| Julian Bakarich (First Respondent) | Sole director of TDC; co-director of TDA; architect and controlling mind directing staff in marketing and executing claims2 | 12 contraventions of s 290-50(1) across 2015–2016 income years4 | Greater of 5,000 penalty units (AU$900,000 at AU$180 per unit) or twice the consideration received/receivable per contravention; aggregate maximum AU$10.8 million3 | AU$4,500,000[cite: 4, 6] | Central to the creation, operation and marketing of each scheme; admitted tax evasion and complete disregard for whether claims were reasonably arguable; significant consideration; penalties, interest and non-financial harm suffered by participants, including the liquidation of Precision Air Pty Ltd; small discounts for late cooperation and the possible Precision Air course of conduct3 |
| The Dream Consortium Pty Ltd (in liq) (Second Respondent) | Primary corporate marketing and contracting vehicle; received contingency fee commissions4 | 12 contraventions of s 290-50(1) across 2015–2016 income years4 | Greater of 25,000 penalty units (AU$4.5 million at AU$180 per unit) or twice the consideration per contravention; aggregate maximum AU$54 million3 | AU$4,500,000[cite: 4, 6] | Direct corporate vehicle for client solicitation and fee extraction; liquidation does not preclude imposition of substantial penalties to serve general deterrence; the Commissioner does not seek leave to enforce the penalty3 |
| The Dream Accountants Pty Ltd (in liq) (Third Respondent) | Corporate accounting vehicle used to prepare inflated R&D schedules and lodge tax returns1 | 12 contraventions of s 290-50(1) across 2015–2016 income years3 | Greater of 25,000 penalty units (AU$4.5 million at AU$180 per unit) or twice the consideration per contravention; aggregate maximum AU$54 million3 | AU$4,500,000[cite: 4, 6] | Same considerations as for Mr Bakarich, whose acts were in substance the acts of TDA; no cooperation from TDA or its liquidators; liquidation does not prevent a penalty imposed for general deterrence3 |
| Thi Cam Tu Nguyen (Fourth Respondent) | Co-director and secretary of TDA; registered tax agent whose details were recorded on AusIndustry applications and who reviewed claim information and prepared R&D Schedules3 | 3 contraventions of s 290-50(1)4 | Greater of 5,000 penalty units (AU$900,000 per contravention) or twice the consideration per contravention; aggregate maximum AU$2.7 million3 | AU$100,000[cite: 4, 6] | Important but relatively confined role; contraventions not deliberate but no basic and prudent checks performed; 30% discount for cooperation and assistance (including an undertaking not to register or practise as a tax agent without the Commissioner’s consent); remorse and contrition; accepted capacity to pay the agreed penalty3 |
For contemporary Australian practitioners, the deterrence principles articulated in Bakarich (No 2) carry heightened prospective force following the enactment of the Treasury Laws Amendment (Tax Accountability and Fairness) Act 2024 (Cth), which took effect from 1 July 202411. Those legislative reforms expanded the ATO’s limitation period to commence promoter penalty proceedings from four years to six years, broadened the nexus requirement from receiving “consideration” to receiving any direct or indirect “benefit” (including intangible or disguised benefits), and increased the statutory maximum civil penalty under s 290-50(4) for bodies corporate and Significant Global Entities (SGEs) to the greatest of 50,000 penalty units, three times the benefits received or receivable, or 10% of the entity’s aggregated turnover up to 2.5 million penalty units—equating to a maximum penalty of approximately AU$780 million8.
The classification of the 12 R&D claims in Bakarich (No 2) as unlawful “tax exploitation schemes” under s 290-65 of Schedule 1 to the TAA 1953 rested on a foundational legal conclusion: it was not reasonably arguable that the claimed R&D tax offsets were available at law under Division 355 of the ITAA 19973. Although the judgment itself did not examine the technical content of individual claims (the contraventions having been admitted or found on the agreed facts), the admitted conduct described in the judgment, read together with the joint ATO–DISR investigation’s finding that clients were encouraged to lodge over-inflated, inaccurate or unsubstantiated claims, points to six interconnected technical, evidentiary, financial, and governance failures that destroyed the legal validity of the underlying claims and exposed the advisors to promoter penalties3.
Under s 355-25(1) of the ITAA 1997, the statutory gateway to the R&DTI is the existence of at least one “core R&D activity”7. The legislation imposes a strict, conjunctive two-limb test7. First, under s 355-25(1)(a), the activity must be an experimental activity whose outcome cannot be known or determined in advance on the basis of current knowledge, information, or experience, but can only be determined by applying a systematic progression of work that is based on principles of established science and proceeds from hypothesis to experiment, observation, and evaluation, leading to logical conclusions7. Second, under s 355-25(1)(b), the experimental activity must be conducted for the purpose of generating new knowledge (including new knowledge in the form of new or improved materials, products, devices, processes, or services)7.
In the schemes promoted by Mr Bakarich, TDC, and TDA—including the claims lodged for clients such as Basebuild Pty Ltd and Precision Air Electrical Pty Ltd—it was admitted that it was not reasonably arguable that the claims were available at law, although the judgment does not describe the particular activities registered for each client3. The admitted facts show a complete disregard as to whether clients’ activities even qualified for the R&D tax incentive, with claims prepared by employees whom Mr Bakarich had recruited with limited or no work experience in R&D tax incentives3. Rather than identifying genuine experimental activities, TDC built registration applications from an online questionnaire asking clients how their company was being innovative in its business methods—an inquiry directed at business innovation rather than at scientific or technological uncertainty3. Where an outcome can be deduced in advance through standard engineering calculations, commercial trial-and-error, or routine configuration of existing equipment, no technological uncertainty exists at law, and the activity is statutorily incapable of qualifying as a core R&D activity7.
Closely linked to the absence of technological uncertainty was the complete failure to demonstrate a “systematic progression of work” proceeding from a falsifiable scientific or technical hypothesis7. In Australian R&DTI law, a compliant hypothesis must articulate a specific causal proposition linking technical variables to a measurable scientific or engineering outcome, formulated prior to the commencement of the experimental activity7.
The mechanics described in Bakarich (No 2) reveal no process directed at this statutory mandate3. Rather than capturing pre-existing technical hypotheses formulated by engineers or scientists during the relevant income year, TDC asked clients to complete an online questionnaire about their operations, finances and how their company was being innovative in its business methods, after Mr Bakarich or other TDC employees had already advised them at an initial meeting that they were eligible3. TDC then used the questionnaire responses to prepare the applications for registration of R&D activities with AusIndustry3. The judgment records no step in this process directed at identifying a technical hypothesis, designing experiments, or conducting systematic observation and evaluation, consistent with the admitted complete disregard as to whether clients’ activities qualified3. The judgment does not identify which, if any, of the claimed activities fell within the statutory exclusions in s 355-25(2) of the ITAA 1997 (such as market research or management studies) or whether any were claimed as “supporting R&D activities” under s 355-30 of the ITAA 1997, so no finding on those provisions can be attributed to the decision3.
Australia’s R&DTI operates within a self-assessment regime, which places the legal burden of proof squarely on the taxpayer (under s 14ZZK of the TAA 1953 in tribunal reviews) to substantiate that registered activities were actually conducted and met statutory definitions4. While Division 355 does not prescribe a single mandatory template for technical records, s 262A of the Income Tax Assessment Act 1936 (Cth) (ITAA 1936) and general substantiation principles require taxpayers to maintain contemporaneous records generated in real time during the execution of the R&D activities7.
In Bakarich (No 2), the R&D Schedules were prepared from questionnaire responses and estimate-based apportionments of profit and loss line items rather than from verified technical or financial records3. The judgment does not describe what technical documentation, if any, the underlying clients held, but no step in the admitted process involved obtaining or reviewing literature reviews, design specifications, test logs, or experimental data3. Relying solely on generic post-hoc questionnaires prepared by an external consultant is evidentially worthless in an ATO or DISR compliance examination4. Justice Kennett made no specific finding about substantiating records, but recorded that Mr Bakarich admitted he took no steps to avoid the contraventions and, when TDA’s R&D Tax Manager advised him that a claim under review was unjustified and should be withdrawn, instructed him to defend it; likewise, Ms Nguyen, the tax agent responsible for ensuring the accuracy of the R&D Schedules, did not conduct basic and prudent checks and controls3.
Even where a taxpayer conducts valid R&D activities, notional deductions under s 355-205(1) of the ITAA 1997 are strictly confined to expenditure incurred during the income year to the extent that the expenditure is incurred on one or more registered R&D activities4. This statutory language (“to the extent”) mandates a rigorous, verifiable nexus between each line item of claimed expenditure and the specific experimental or supporting activities, alongside a defensible, evidence-based apportionment methodology where staff or overheads straddle both R&D and ordinary commercial operations7.
The joint ATO–DISR investigation found that clients were encouraged to lodge over-inflated, inaccurate or unsubstantiated R&DTI claims, and the judgment identifies the mechanism by which expenditure was attributed to R&D6. The first defect was the method of attribution itself: TDA prepared each R&D Schedule by apportioning line items in the taxpayer’s profit and loss statement according to an estimate of how much those items were devoted to R&D activities3. Because TDC’s fee was a percentage of the refund, Mr Bakarich encouraged employees to maximise claims to secure higher commissions and instructed them to maximise the value of claims regardless of their legitimacy3. The judgment does not refer to any timesheets, cost-allocation drivers, or other objective basis for the estimated apportionments3. The second defect, identified by the ATO investigation rather than by the Court, was that claims were unsubstantiated6. ATO compliance action across the approximately 138 associated taxpayers resulted in over AU$14.5 million in tax shortfalls, administrative penalties and interest charges6. By estimating apportionment percentages with the aim of maximising the refundable offset, TDC and TDA severed the statutory nexus required by s 355-2053.
A critical commercial feature examined in Bakarich (No 2)—and repeatedly flagged by the ATO as a primary compliance risk indicator—was TDC’s contractual fee structure4. TDC contracted with clients on a contingency basis, charging a percentage commission (typically 15% of the refund received by the taxpayer, with approximately AU$590,808 receivable by TDC across the 12 prosecuted schemes, of which approximately AU$438,692 was actually received) tied directly to the R&D tax refund received by the client3.
From a legal and governance perspective, this contractual architecture had two profound consequences4. First, under s 290-60(1)(b) of Schedule 1 to the TAA 1953, the receipt of percentage-based fees directly satisfied the statutory requirement that the promoter (or an associate) receive “consideration in respect of that marketing or encouragement”4. Second, as the admitted facts recorded by Justice Kennett show, the contingency fee created a perverse structural incentive: Mr Bakarich encouraged employees to maximise R&D claims to secure higher commissions for TDC, so that the larger the claim, the larger the fee3. Crucially, the commercial risk was entirely asymmetrical, a feature Justice Kennett identified as a reason why general deterrence is important for promoters who profit while participants may be subject to penalties3. When the ATO reviewed the claims, scheme participants bore administrative penalties and interest charges and, in the case of Precision Air Pty Ltd, liquidation, whereas TDC had already received approximately AU$438,692 of its fees before TDC and TDA entered external administration on 27 January 20173.
The final systemic failure elucidated in Bakarich (No 2) involved the abdication of professional gatekeeping duties under the Tax Agent Services Act 2009 (Cth) (TASA) and Division 290 of the TAA 19531. Under s 290-60(2) of Schedule 1 to the TAA 1953, an entity is not a promoter merely because the entity provides advice about a scheme12. However, this “mere advice” safe harbour is lost the moment a tax practitioner steps beyond independent, objective legal assessment and actively facilitates, rubber-stamps, or executes the implementation of an unarguable scheme4.
In Bakarich (No 2), Mr Bakarich was not a registered tax agent; obtaining advice from a registered tax agent was a prerequisite for the registration of the R&D claims, and Ms Nguyen became registered as a tax agent in the course of establishing the accounting practice that traded as TDA3. Ms Nguyen, as co-director of TDA and a registered tax agent, reviewed the AusIndustry applications, had her details recorded in the tax agent section of the forms, reviewed the information forming the clients’ claims and prepared R&D Schedules, yet the applications indicated that she had provided advice that was not actually provided to the clients, and she did not conduct basic and prudent checks and controls3. Although the Commissioner did not allege that the schemes as they related to Ms Nguyen involved tax evasion or that she knew of false representations, and the parties accepted that her contraventions were not deliberate (she relied on Mr Bakarich’s explanation of the R&D regime and accepted that she should have made further enquiries), Ms Nguyen admitted contravening s 290-50(1) in respect of the 3 schemes, and the Federal Court imposed the agreed AU$100,000 civil penalty, which reflected a 30% discount for her cooperation and assistance3.
The following table maps each governing statutory requirement across Division 290 of Schedule 1 to the TAA 1953 and Division 355 of the ITAA 1997 against the specific factual and evidentiary failures identified in Commissioner of Taxation v Bakarich (No 2) [2024] FCA 1448 and the corresponding legal standards; because the contraventions were admitted (or found against TDC and TDA on the agreed facts), the Federal Court did not itself apply the Division 355 standards to the individual claims3.
| Statutory Provision & Legal Requirement | Taxpayer / Promoter Failure in Bakarich (No 2) | Federal Court Standard of Proof & Evidentiary Threshold |
|---|---|---|
| Promoter Status Section 290-60(1) of Schedule 1 to the TAA 1953 Prohibits marketing or encouraging a scheme for consideration while playing a substantial role1. | TDC actively solicited clients via marketing materials, word of mouth and referral partners, charging contingency fees based on a percentage of R&D refunds3. Mr Bakarich exercised total operational control over employees preparing the claims4. Ms Nguyen’s tax agent details were recorded on the AusIndustry applications, which indicated she had provided advice that was not actually provided to clients, and she did not conduct basic and prudent checks3. | Agreed Facts (s 191 Evidence Act 1995 (Cth)): The contraventions were established by statements of agreed facts and admissions; findings against TDC and TDA were made in the terms of the Bakarich statement of agreed facts because they chose not to appear or contradict it3. The “mere advice” exception (s 290-60(2)) fails where an advisor actively prepares, packages, and lodges unverified claims for contingent consideration4. |
| Tax Exploitation Scheme & Reasonably Arguable Position (RAP) Sections 290-65(1) & 284-15 of Schedule 1 to the TAA 1953 Dominant purpose of obtaining a scheme benefit that is not reasonably arguable at law2. | 12 schemes involving 11 taxpayers were promoted to obtain refundable R&D tax offsets (around AU$7m had they gone undetected), and it was admitted that the claims were not reasonably arguable3. The success of the schemes relied on the Commissioner never discovering that the taxpayers were not entitled to the offsets, as reflected in Mr Bakarich’s admission of tax evasion3. | Objective Legal Threshold: Under s 284-15, a position is reasonably arguable only if, having regard to relevant statutory and judicial authorities, it is about as likely to be correct as incorrect (50/50 threshold)2. Complete absence of factual or technical foundation renders a claim unarguable ab initio4. |
| Statutory Exceptions & Due Diligence Defence Section 290-55 of Schedule 1 to the TAA 1953 Exemption for reasonable mistake of fact or taking reasonable precautions and exercising due diligence12. | Clients completed an online questionnaire about their operations, finances and innovation, and Mr Bakarich admitted that neither he nor any TDC or TDA employee took any steps to avoid the contraventions3. Ms Nguyen reviewed the applications and prepared R&D Schedules but did not conduct basic and prudent checks and controls3. | Not Relied Upon: No exception under s 290-55 was raised; the Court took into account that Mr Bakarich and Ms Nguyen took no steps to avoid the contraventions (s 290-50(5)(g)), and noted that, because of Mr Bakarich’s admitted tax evasion, s 290-55(6) displaced the time limit in s 290-55(4)3. |
| Core R&D Activities: Technological Uncertainty & New Knowledge Section 355-25(1)(a) & (b) of the ITAA 1997 Outcome cannot be known in advance based on current knowledge/experience; conducted to generate new knowledge7. | The judgment does not describe the activities registered for individual clients (e.g., Basebuild Pty Ltd, Precision Air Electrical Pty Ltd); it records that the claims were admittedly not reasonably arguable and that there was complete disregard as to whether clients’ activities even qualified3. | Objective Competency Standard: Requires contemporaneous evidence establishing the industry “state of the art” (via prior art searches and literature reviews) and proving why a competent professional in the field could not deduce the technical outcome without experimentation7. |
| Core R&D Activities: Systematic Progression of Work Section 355-25(1)(a)(i)–(ii) of the ITAA 1997 Must be based on established science and proceed from hypothesis to experiment, observation, evaluation, and logical conclusions7. | The judgment records no hypothesis-testing or experimental process3. Registration applications were prepared by TDC from online questionnaire responses after clients had already been told they were eligible3. | Contemporaneous Scientific Method: Requires real-time documentation of a specific, falsifiable causal hypothesis prior to testing, coupled with structured experimental designs, test logs, quantitative data observation, and documented logical conclusions7. |
| Supporting R&D Activities & Statutory Exclusions Sections 355-25(2) & 355-30 of the ITAA 1997 Activities directly related to core R&D (and for excluded or production activities, undertaken for the dominant purpose of supporting core R&D)2. | The judgment does not address supporting R&D activities or the s 355-25(2) exclusions; it records only that the claims were not reasonably arguable3. | Nexus & Dominant Purpose Test: If no valid core R&D activity is proven, all supporting activities automatically fail2. For production or s 355-25(2) excluded activities, the taxpayer must prove via contemporaneous project charters that supporting the core experiment was the ruling, prevailing, or most influential purpose2. |
| Notional R&D Deductions & Expenditure Apportionment Sections 355-100 & 355-205 of the ITAA 1997 Expenditure must be actually incurred during the income year to the extent it is on registered R&D activities4. | TDA prepared R&D Schedules by apportioning profit and loss line items according to an estimate of R&D use, and employees were instructed to maximise claims regardless of legitimacy, increasing TDC’s percentage-based fee3. | Verifiable Financial Nexus: Requires source accounting proof that expenditure was incurred (invoices, general ledger, payroll) plus granular, contemporaneous time-tracking (e.g., activity-based timesheets) and objective apportionment methodologies separating R&D from BAU costs7. |
To assist Australian corporate taxpayers, in-house tax counsel, and R&DTI practitioners in translating the judicial warnings of Commissioner of Taxation v Bakarich (No 2) [2024] FCA 1448 into robust compliance practices, the following two applied industry case studies model realistic, hypothetical commercial scenarios that illustrate technical, financial, and contractual pitfalls of the kind exposed in the litigation, together with other common R&DTI compliance risks3. Each case study provides a rigorous, step-by-step remedial blueprint demonstrating how to restructure project governance, technical documentation, contractual engagement terms, and financial tracking to satisfy Division 355 of the ITAA 1997 and eliminate Division 290 promoter and shortfall penalty exposure4.
Consider Southern Cross Climate & Electrical Pty Ltd (SCCE), an Australian mechanical and electrical engineering contractor with an aggregated turnover of AU$18 million, below the AU$20 million refundable R&DTI threshold under s 355-100 of the ITAA 19974. SCCE was engaged under a fixed-price commercial design-and-construct subcontract to install a high-efficiency Heating, Ventilation, and Air Conditioning (HVAC) and cleanroom pressurisation system for a pharmaceutical manufacturing facility in western Sydney3. After the close of the financial year, SCCE is approached by an outbound R&D consultancy operating on a “No Refund, No Fee” 30% contingency commission4. The consultant provides SCCE’s Managing Director with a two-page generic questionnaire4. Based on brief responses stating that the cleanroom HVAC installation was complex, custom-engineered, and required multiple balancing adjustments on site, the consultant lodges an AusIndustry registration for a purported Core R&D Activity titled “Development of Novel Pharmaceutical Cleanroom HVAC Pressurisation System” and instructs SCCE’s accountant to lodge an ATO R&D tax schedule claiming AU$1.8 million in notional R&D deductions to generate a refundable offset of AU$783,0004.
This baseline posture fails catastrophically under the legal principles enforced in Bakarich (No 2) for three reasons4. First, it commits the whole-of-project business-as-usual (BAU) fallacy: installing and balancing commercial HVAC equipment to meet a client’s cleanroom specification is routine commercial engineering, and claiming the entire commercial fit-out reflects the same disregard for whether activities genuinely qualify that underpinned the schemes in Bakarich (No 2), in which scheme participants such as Precision Air Electrical Pty Ltd incurred administrative penalties and interest charges and Precision Air Pty Ltd was placed in liquidation3. Second, the claimed work lacks technological uncertainty under s 355-25(1)(a), as standard psychrometric calculations and proportional-integral-derivative (PID) damper controls represent established engineering knowledge, while site balancing constitutes routine commissioning excluded under s 355-25(2)(f) (quality control and routine testing)7. Third, apportioning 80% of commercial project labour and standard plant equipment without contemporaneous timesheets or plant exclusions violates s 355-205 and exposes SCCE to 50% to 75% administrative culpability penalties for recklessness or intentional disregard of the taxation law4.
To insulate both SCCE and its external tax advisors from Division 290 promoter scrutiny and ATO high-risk audit triggers, SCCE must immediately terminate the percentage-of-refund engagement8. SCCE executes a formal fixed-fee or time-billed engagement letter with a registered tax agent and R&DTI specialist whose scope is strictly defined as providing independent legal and technical eligibility verification under the Tax Agent Services Act 2009 (Cth)4. The engagement letter explicitly prohibits success fees, bonus structures, or commissions tied to claim quantum, thereby removing the consideration nexus under s 290-60(1)(b) and preserving the “mere advice” protection under s 290-60(2) of Schedule 1 to the TAA 19534. Simultaneously, SCCE establishes an internal R&D Governance Committee comprising its Chief Engineer, Chief Financial Officer (CFO), and External Registered Tax Agent, mandating that no activity may be registered with AusIndustry unless signed off via a formal pre-registration Technical and Financial Verification Protocol4.
The R&D Governance Committee conducts a rigorous technical deconstruction of the pharmaceutical cleanroom project to excise all BAU design, procurement, ductwork fabrication, and site commissioning8. Through engineering interviews, the Committee isolates one specific, genuine technical bottleneck encountered during the project7. Specifically, the pharmaceutical facility required ultra-rapid pressure recovery (under 1.5 seconds) across a cascade of four ISO Class 5 airlocks subject to high-frequency automated guided vehicle (AGV) door openings, without inducing laminar airflow turbulence above a ±20% velocity variance7. Existing commercial variable air volume (VAV) venturi valves and standard PID control loops suffered from mechanical hysteresis and acoustic resonance, causing pressure oscillations that breached containment standards7. Prior to commencing experimental work, SCCE’s Chief Engineer documents a comprehensive literature and patent review (searching ASHRAE transactions, IEEE control systems databases, and OEM technical manuals from Siemens, Trox, and Honeywell)7. This documented prior-art search proves that no existing commercial HVAC damper actuator or published control algorithm could achieve sub-1.5-second pressure stabilisation across multi-room cascades without disrupting laminar flow uniformity, thereby establishing that the outcome could not be known or determined in advance under s 355-25(1)(a)7.
Instead of registering the commercial HVAC installation, SCCE scopes a narrow, highly defensible R&D project strictly confined to developing a novel predictive fluid-dynamic control actuator and algorithm7. Before building the test rig, the engineering team records a formal, falsifiable hypothesis in its Project Initiation Charter: “If a dual-stage piezoelectric micro-baffle array is coupled with a feed-forward Kalman filter algorithm utilizing upstream differential pressure derivatives (dP/dt) sampled at 1 kHz, then pressure cascade recovery across coupled ISO Class 5 airlocks during rapid door actuation will stabilise within 1.5 seconds while maintaining laminar velocity variance within ±20%, overcoming the mechanical latency of conventional VAV venturi valves”7.
To test this hypothesis systematically, SCCE constructs a dedicated, off-site modular test plenum at its warehouse—completely segregated from the client’s commercial construction site—and executes four structured experimental phases7. In the first phase, the team conducts baseline benchmarking and variable isolation by testing conventional venturi valves under simulated AGV door openings to quantify exact latency and turbulence decay curves7. In the second phase, the team fabricates and tests three distinct piezoelectric micro-baffle geometries across Reynolds numbers ranging from 2,000 to 20,000, measuring acoustic resonance and pressure drop7. In the third phase, engineers execute 45 controlled bench trials varying sensor sampling frequencies and derivative gain matrices, logging real-time differential pressure telemetry (ΔP, in pascals) and hot-wire anemometer velocity profiles (in m/s)7. In the fourth phase, the team performs failure analysis and formulates logical conclusions, documenting why Geometry B failed due to vortex shedding at high differential pressures and proving why Geometry C paired with a modified Extended Kalman Filter satisfied the hypothesis parameters7.
To satisfy DISR and ATO evidentiary standards and avoid the Bakarich post-hoc questionnaire trap, SCCE compiles an immutable contemporaneous evidence pack in real time during the income year4. This dossier preserves dated literature review PDFs and OEM correspondence confirming the physical limitations of off-the-shelf VAV valves, version-controlled CAD schematics and Computational Fluid Dynamics (CFD) simulation files for the three experimental baffle prototypes, raw time-stamped CSV telemetry logs exported directly from the test plenum’s differential pressure transducers and anemometers across all 45 experimental runs, and fortnightly Engineering R&D Review Minutes signed by the Chief Engineer evaluating failed runs and authorizing parameter modifications7.
Because the technical problem was identified during a commercial subcontract for a pharmaceutical client, SCCE must ensure its R&D claim satisfies the “conducted for the R&D entity” (on own behalf) requirement in s 355-210 of the ITAA 1997, which prevents double-dipping or claiming funded research performed as a contractor for another party4. SCCE executes a formal intellectual property (IP) and scope addendum to its commercial subcontract with the pharmaceutical principal. The contract addendum explicitly stipulates that SCCE is contracted solely to deliver a functional cleanroom meeting performance specifications for a fixed commercial price, that the pharmaceutical client is not commissioning or reimbursing SCCE to conduct R&D, that SCCE bears 100% of the financial and technical risk of its off-site experimental baffle and algorithm development, and that all foreground IP, patents, and control code generated from the experimental work vest exclusively in SCCE for future commercialisation across its broader client base4.
Finally, SCCE overhauls its cost-accounting architecture to eliminate any BAU contamination7. The CFO creates a dedicated R&D Cost Centre (CC-901: Predictive Airlock Baffle R&D) in the company’s ERP general ledger, completely segregated from the commercial job code (JOB-402: Pharma Cleanroom Fit-out)7. Only the three mechanical and control engineers working on the off-site test plenum log hours to CC-901, completing daily activity-based timesheets that link each hour to a specific registered Core (CORE-01) or Supporting (SUPP-01: Test Rig Fabrication) activity, countersigned weekly by the Chief Engineer7. All site labour, commercial chillers, standard ductwork, and commissioning costs on the pharmaceutical client’s premises remain in JOB-402 and are excluded from the R&D tax schedule8. Consequently, SCCE lodges a watertight, fully substantiated R&D expenditure claim of AU$245,000 (down from the inflated baseline claim of AU$1.8 million), securing a legally unimpeachable refundable offset of AU$106,575 with zero exposure to shortfall or promoter penalties4.
Consider LogiChain AI Pty Ltd (LogiChain), a Melbourne-based logistics software company developing a cloud-based fleet routing and warehouse management platform7. LogiChain engages a boutique R&D promoter who promises to unlock maximum cash refunds for software development under a percentage-based success fee1. Following a brief onboarding questionnaire, the promoter registers the entire software development lifecycle—including user interface (UI/UX) design, database migration to Amazon Web Services (AWS), API integrations with third-party ERP systems, and the deployment of an off-the-shelf open-source machine learning library (XGBoost) for delivery time estimation—as a single Core R&D Activity4. Furthermore, to inflate the expenditure base, the promoter advises LogiChain’s founder to interpose a related service entity (LogiChain Dev Trust, controlled by the founder) that invoices LogiChain AU$1.5 million in “R&D management fees” (carrying a profit markup and left unpaid at year-end), inflating LogiChain’s total claimed R&D expenditure to AU$2.2 million on the ATO R&D tax schedule4.
This arrangement violates both Division 355 of the ITAA 1997 and the promoter penalty prohibitions in Division 290 of Schedule 1 to the TAA 19534. First, integrating existing APIs, migrating cloud databases, building user interfaces, and training standard open-source machine learning models on company data involve applying established computer science methodologies where technical outcomes are known in advance7. Moreover, software developed for internal administration is explicitly excluded from core R&D under s 355-25(2)(h) of the ITAA 19977. Second, artificially inflating R&D deductions through associated entity arrangements without commercial substance directly triggers ATO Taxpayer Alert TA 2023/4, breaches the associate payment rules in s 355-480 of the ITAA 1997 (which disallow notional deductions for expenditure incurred to an associate until the amount is actually paid), and violates the arm’s-length rules in s 355-400 (which strip out non-arm’s-length markups)8. Promoting such an arrangement without a reasonably arguable basis constitutes a classic tax exploitation scheme under s 290-65 of Schedule 1 to the TAA 19534.
To eliminate the Taxpayer Alert TA 2023/4 and Division 290 tax exploitation scheme risks, LogiChain’s Board immediately terminates the contingency promoter and retains a reputable, fixed-fee registered tax agent and software R&D specialist4. LogiChain cancels the uncommercial management fee invoices issued by LogiChain Dev Trust8. Where software engineers are genuinely employed by a related group entity, intercompany recharges are restructured strictly at actual cost—eliminating any profit markup in compliance with the arm’s-length rules in s 355-400—and are settled via actual cash payment prior to 30 June of the relevant income year to satisfy the strict associate payment requirement under s 355-480 of the ITAA 19978.
Aligned with DISR’s published guidance on artificial intelligence and machine learning RDTI eligibility, LogiChain’s Chief Technology Officer (CTO) and external technical advisor audit the software roadmap to separate routine software engineering from genuine computer science research7. All UI/UX front-end work, AWS cloud configuration, CRUD database architecture, third-party ERP API connectors, and routine hyperparameter tuning of standard XGBoost models are classified as ineligible BAU software development and excised from the claim7.
The CTO then isolates one specific algorithmic research stream where current computer science knowledge was demonstrably insufficient: developing a real-time, decentralized Spatio-Temporal Graph Neural Network (ST-GNN) capable of executing dynamic multi-depot vehicle re-routing on edge-computing telematics hardware with strictly bounded memory (≤ 512 MB) and sub-50-millisecond inference latency under stochastic traffic disruptions7. Prior to coding the experimental architecture, the data science team compiles a formal Prior Art and Literature Review analysing recent peer-reviewed publications from NeurIPS, ICML, and IEEE Transactions on Intelligent Transportation Systems7. The review demonstrates that existing ST-GNN architectures suffer from quadratic memory complexity O(N²) as graph nodes scale, causing out-of-memory crashes on edge hardware, and that standard model quantization degrades routing accuracy below commercially viable thresholds7. Consequently, whether a sparse attention graph pruning mechanism could preserve routing optimality within a 512 MB edge memory footprint could not be known or determined in advance under s 355-25(1)(a)7.
LogiChain embeds the statutory requirements of s 355-25(1)(a) directly into its software engineering workflow so that technical evidence is generated automatically and contemporaneously7. In Confluence, the Lead Data Scientist creates an R&D Experiment Charter prior to Sprint 1 recording the formal hypothesis: “Hypothesis H1: Implementing a dynamic top-k topological edge-pruning layer combined with mixed-precision 4-bit quantization within a Spatio-Temporal Graph Neural Network will reduce memory complexity from O(N²) to O(N·k), achieving inference latency below 50 ms on ARM Cortex-A72 edge hardware for graphs up to 10,000 nodes while maintaining optimality gap degradation below 2% compared to exact Mixed-Integer Linear Programming (MILP) solvers”7.
Using MLflow and Git, the team tracks 28 distinct experimental training and edge-deployment runs over four months7. Each run logs the exact mathematical modifications to the loss function and graph convolution layers, hardware memory consumption curves, inference latency distributions, and optimality gaps against benchmark datasets7. Failed experiments—such as Run 12, where uniform 4-bit quantization caused gradient vanishing across temporal edges—are explicitly preserved in Confluence with post-mortem evaluations leading to the next logical iteration7.
Under s 355-30 of the ITAA 1997, activities that are not experimental in themselves can only be claimed if they are “directly related” to a Core R&D Activity, and where they fall within a s 355-25(2) exclusion or produce goods or services, they must be undertaken for the dominant purpose of supporting the core experiment2. Applying this strict test, LogiChain claims only two narrow supporting activities: writing a synthetic stochastic traffic disruption generator specifically required to feed test data into the experimental ST-GNN (SUPP-01), and configuring the hardware-in-the-loop ARM Cortex-A72 bench test environment (SUPP-02)2. All data cleaning for commercial clients and commercial platform deployment are excluded7.
To satisfy the ATO’s substantiation requirements under s 355-205 and defeat any allegation of arbitrary expense apportionment, every Jira ticket in LogiChain’s engineering sprints is assigned a mandatory taxonomy field vetted by the CTO: RD-CORE-01, RD-SUPP-01, RD-SUPP-02, or BAU-COMMERCIAL7. Engineers log hours directly against Jira tickets via Tempo Timesheets as work is performed8. At year-end, the CFO exports the audited Tempo timesheet ledger to calculate exact staff apportionment ratios8. For example, where a Senior Data Scientist spends 1,140 verified hours on RD-CORE-01 out of 1,900 total annual working hours (60%), exactly 60% of her gross salary, superannuation, and payroll tax is allocated to the R&D schedule7. Similarly, cloud GPU compute costs (AWS EC2 P4d instances) are claimed strictly using AWS Cost Explorer tags linked exclusively to the experimental MLflow training environment, completely segregated from the company’s commercial production servers7.
| Governance & Compliance Dimension | Defective Bakarich Practice (High Risk of Disallowance & Penalties) | Remedial Best-Practice Standard (Fully Compliant with Div 355 & Div 290) |
|---|---|---|
| Advisor Engagement & Fee Model | Contingency fee / percentage of R&D tax refund (typically 15% in Bakarich); an unregistered business coach directing claims3. | Fixed-fee or hourly time-billed professional engagement with a registered tax agent; zero financial link to refund quantum8. |
| Technical Scoping & Registration | Online questionnaires about operations, finances and business innovation; eligibility assurances given before any technical review; registering activities without regard to whether they qualified3. | Pre-experiment technical charters; literature/patent searches proving state-of-the-art gaps; isolating specific scientific/technical uncertainties7. |
| Experimentation & Evidence | Registration applications prepared by staff with limited or no R&D tax experience from questionnaire responses; absence of hypotheses, test protocols, or failure logs3. | Contemporaneous hypotheses recorded prior to testing; version-controlled test logs, telemetry data, Git/MLflow histories, and failure evaluations7. |
| Contractual & Related-Party Terms | Commercial contracts where clients pay for the work; inflated related-party management fees unpaid at year-end (illustrative risks drawn from the case studies above rather than findings in Bakarich (No 2))8. | Explicit IP and financial risk carve-outs satisfying s 355-210; associate charges strictly at cost (s 355-400) and paid in cash before 30 June (s 355-480)8. |
| Financial Nexus & Apportionment | Estimate-based apportionment of profit and loss line items to R&D, maximised regardless of legitimacy to increase the fee3. | Dedicated general ledger cost centres; daily activity-based timesheets linked to specific Core/Supporting codes; direct tagging of R&D consumables/compute7. |
The Federal Court’s decision in Commissioner of Taxation v Bakarich (No 2) [2024] FCA 1448—imposing AU$13.6 million in civil penalties on the promoters, following ATO compliance action that resulted in over AU$14.5 million in tax shortfalls, administrative penalties and interest charges across the approximately 138 associated taxpayers—marks a decisive inflection point in the administration of the Australian R&D Tax Incentive6. Viewed alongside the post-1 July 2024 expansion of Division 290 maximum penalties to AU$780 million and the introduction of shortfall interest charges on overclaimed refundable R&D offsets from 1 April 2025, the judgment sends an unmistakable message to corporate boards and the tax profession7.
First, the self-assessment nature of initial AusIndustry registration is not an endorsement of eligibility4. Registration under s 27A of the IR&D Act is an administrative prerequisite, not a substantive shield against subsequent DISR formal examination or ATO financial audit4. Second, corporate directors cannot outsource their statutory tax compliance responsibilities to aggressive contingency-fee promoters; when an unarguable R&DTI claim collapses under audit, the corporate taxpayer bears the full brunt of the clawback, interest, and shortfall penalties8. Finally, registered tax agents and corporate advisors must recognize that lending professional credentials to unverified, template-driven R&D claims destroys the statutory “mere advice” protection under s 290-60(2) and exposes both the practitioner and their corporate entities to catastrophic personal and corporate liability1. Only by embedding rigorous, contemporaneous scientific governance, objective prior-art verification, and granular cost-accounting nexus into the R&D lifecycle can Australian enterprises securely access the statutory benefits of Division 355 of the ITAA 19974.